I really like the idea of LR being "anti-photoshop". I think it is an interesting exercise to ponder what features would make sense to include in LR without it loosing its "soul".
And that's exactly what Mark Hamburg was going for when he designed Lightroom...in truth really, he thought of LR ad the "Un-Photoshop" sort like 7-UP being the Un-Cola, not really the anti-Photoshop :~)
The main fulcrum in the decision axis is what image processing is
BEST done with raw processing and what is best done in Photoshop? Since LR is an extension of Camera Raw's processing pipeline and that pipeline is designed and controlled by Thomas Knoll (who also happened to develop Photoshop), the push to add features and functionality to Camera Raw is really based on what is the optimal processing for raw files. Those things that can (and should) be done in ACR are done in ACR and the rest deferred to Photoshop.
There is no conscience effort to keep stuff out of ACR/LR to preserve Photoshop sales...seriously, if you think that, you don't understand the Photoshop marketplace. Photographers make up a very, very small % of the Photoshop user base (under 10% and getting smaller). The real line in the sand is what is best done with a metadata editor and what is best done with a pixel editor.
There's a lot of editing that's really easy to do (easy meaning engineering-wise) with pixels that is very, very difficult to do with metadata edits. If you understand the distinction, it becomes easier to understand what to hope for and ask for in terms of new features and functionality in ACR/LR. And, remember, both pipelines are tied together...what goes in one, will be in the other. I'm sure there will be some nice surprises showing up in LR 5 and ACR 8...(in fact, I already know some of them).
But...it's not really useful asking for features that are conceptually not designed for the basic task that ACR/LR was/is designed for; raw image processing.